Definición de la integración

An alternative laser material that has the potential for scaling up to higher powers is Nd:YLF. Yttrium lithium fluoride (YLF) has the same form as the tetragonal crystal structure scheelite (CaWO4) with yttrium corresponding to

calcium, lithium to tungsten and fluoride to oxygen [2, 3].

Figure 3.1 Tetragonal crystal structure scheelite (CaWO4) [2].

In Nd YLF, Nd+3 substitutes yttrium in the lattice with a concentration of around 1.37x1020 cm-3 which is equivalent to 1% yttrium being replaced by Nd. Because of its crystal structure, Nd:YLF is naturally birefringent and many of its properties are dependent on the polarisation direction through the crystal. Table 3.1 shows the various material properties of Nd:YLF compared to another commonly used laser material Nd:YAG [4]:

Material Property Nd:YAG Nd:YLF

Heat Conductivity at 300K(Wm-1_K-1_{) 13 6}

Expansion Coefficient αT (10-6K-1) 8.2 c:8,a:13

Elastic Modulus (GPa) 310 75

Poisson ratio 0.3 0.33

Refractive index at ~1µm 1.82 π:1.470,σ:1.448

dn/dt at 300K (10-6_K-1_{) 7.3 π:-4.3,σ:-2.0}

Tensile strength (107 _Nm-2_{) 21 3.3}

59 Nd:YLF has the attraction that on its σ-polarisation (corresponding to 1053nm operation in table 3.1) the thermal lensing is very weak. This is due to a combination of a small change in refractive index with temperature dn/dT (~-2.0×10-

6_K-1_{) and a positive contribution to thermal lensing from end-face bulging which}

offsets the negative lensing. The net result is that for the same pumping conditions and hence, amount of heat generation, the thermal lens on the 1053nm transition in Nd:YLF is ~6 times weaker than in Nd:YAG. However, Nd:YLF does suffer from the problem that its has a stress-fracture limit that is ~ 5 times lower than that of Nd:YAG meaning that in previous experiments involving 1% dopant concentration Nd:YLF thermal fracture was found to occur for diode pump power in the range of ~14W of absorbed pump power [2] in a pump waist size of ~100µm. Nd:YLF also has the disadvantage to suffer largely from energy transfer up-conversion, resulting in a significant reduction in the fluorescence lifetime τf and a significant increase in the thermal loading which places extra constraints on the laser.

In Chapter 2, we mentioned the spectroscopic losses that are encountered within Nd: doped laser materials. Several authors [5-12] have measured the Stark- energies of various multiplet levels in Nd:YLF. Although the majority of heat generation at low powers within these materials is due to quantum defect heating, as the incident pump power is increased, other spectroscopic losses, namely energy transfer upconversion, that are present and increase non-linearly with increasing pump power dramatically increasing heat generation and its detrimental effects.

Figure 3.2 shows a simplified energy level diagram for Nd:YLF, displaying fluorescence and the various spectroscopic losses that lead to excess heat generation. Pump radiation (797nm) is absorbed from the ground state (4I9/2) to the pump level

(4F5/2) from where it non-radiatively decays to the upper-laser level. Under non-

lasing conditions, the upper-laser level decays via four fluorescent processes emitting at ~1830nm (4F 3/2→4I15/2), ~1330nm (4F3/2→4I13/2), ~1050nm (4F3/2→4I11/2) and

60

Figure 3.2Diagram showing various energy levels for Nd+3_{, displaying fluorescence and various} spectroscopic loss mechanisms. ESA PR/LR – excited state absorption of pump radiation / laser radiation. ETU – Energy transfer up-conversion. CR – cross relaxation. All dashed lines indicate heat

generating processes[2].

These processes are shown in figure 3.2 by a solid arrow and the following mulitphonon decay by a dotted arrow. In 1998, Pollnau et al [11], utilised the branching ratios B4j from the 4F3/2 level (including radiative and mulitphonon decay)

and the fluorescence wavelengths to calculate the fraction of absorbed pump power

ρnl converted to heat, under non-lasing conditions, without the effect of energy

transfer up-conversion (see later in the Chapter). Assuming that all the fluorescence processes occurred from the lowest Stark level in the 4F3/2 multiplet and using a

pump wavelength of 797nm, a value of ρnl ~ 25% was obtained. This is only a small

increase from the quantum defect heating, under lasing conditions, of ~24% [2]. Cross-relaxation occurs where an excited ion transfers part of its energy to an unexcited ion so that both are in the 4_I

15/2 level. Excited state re-absorption of the

pump (ESAPR) occurs when an excited ion in the 4_F

3/2 multiplet absorbs a pump

photon and is excited up to the 2_D

5/2 multiplet. Excited-state absorption of the laser 4 D5/2 2 P3/2 PUMP CR 1 PUMP PUMP ETU 1 ETU 4 ETU 3 ETU 2 ~1.3 mµ ~1 mµ

Fluoresence Cross-relaxation ESAPR ETU ESALR

61 radiation (ESALR) is when an excited ion in the 4_G

7/2 multiplet (which can be

populated by ETU) absorbs a laser photon and is excited to either the 4D5/2 or the 2_P

3/2 multiplets (this is a very weak effect because the lifetime of the 4G7/2 multiplet

is very fast. One effect that has to be considered in any power-scaling approach on any Nd doped laser is that of energy transfer up-conversion (ETU) [10-16], whereby, one excited ion in the upper laser level, relaxes down to a lower lying levels and transfers its energy to a neighbouring excited ion, also in the 4F3/2 level, which is the

raised (up-converted) to a higher level.

Figure 3.3 shows the absorption spectra for a 1.15% doped Nd: YLF rod taken with a Perkin-Elmer spectrophotometer [2], in the wavelength range of interest for diode pumping. It can be seen that the two highest absorption peaks occur at 792nm and 797nm respectively, for light polarised parallel to the c-axis of the crystal. It must be noted that since a diode pump source has a finite bandwidth (typically ~3nm) that there will be a distribution of absorption coefficients within the laser rod itself. If the diode pump source were to be temperature tuned away from the material absorption peaks, most of the pump light will be absorbed along a longer length of crystal, however the light closer to the absorption peak will be absorbed over a much smaller distance. This in itself can have important consequences when designing a laser with the aim of reducing up-conversion, thermal lensing and the chance of thermal fracture.

Figure 3.3 The measured absorption of Nd: YLF NB. The noise on the two largest absorption peaks is caused by the lack of sensitivity of the detector at the probe light wavelengths.

62

Figure 3.4 The calculated emission cross-section [17] for light polarised parallel to the crystal c-axis (π polarisation) and light perpendicular to the crystal c-axis (σ polarisation).

Figure 3.4 shows the calculated emission cross-sections for the Nd 4F3/2→4I11/2

transition in Nd:YLF [17]. The two lines with the largest cross-section occur at 1053nm (σ polarisation) and 1047nm (π polarisation). The upper-laser lifetime has been measured to be ~485µs [17] which is approximately twice that of Nd: YAG [4]. This feature of Nd: YLF was of particular interest for high power Q-switched operation and amplifier stages because of the longer energy storage time. It has been suggested and recently researched that degradation in the use of Nd: YLF for Q- switched operation [2] is due to energy transfer up-conversion.